"Hugging Face tokens are notorious for allowing access to private AI models," said Berkovich. "The leaked Hugging Face token belonging to an AI 50 company could have exposed access to ~1,000 private ...

Google Threat Intelligence Group has found hackers are now using AI tools like Gemini to make malwares that can self write, ...

The median expert also sees AI advancing more slowly than makers of frontier AI models, which foresee human- or ...

近日，绿盟科技发布《2025年全球云上数据泄露风险分析报告（第八期）》，深入剖析了当年9月至10月期间全球范围内发生的多起云上数据泄露事件。报告指出，随着人工智能技术的广泛应用，云上数据安全面临前所未有的挑战，而配置错误、凭证泄露与供应链攻击依然是数 ...

在当今AI发展的浪潮中，智能体逐渐崭露头角，成为了新一代的平台级机会。随着Mem0成功融资2400万美元，其以“记忆层”为核心的布局，或许将重塑智能体的基础设施。生成式AI的崛起，让我们见证了语言模型在编程、合同分析和图像生成等领域的卓越表现，但这些智能体却普遍存在一个致命的缺陷：无法记忆。

The flaws let attackers abuse masked paths, console bind-mounts, and redirected writes to escape containers and gain host ...

Do you thrive on automation and innovation? Do you enjoy helping a company work more efficiently and effectively? If so, we look forward to receiving your application.

Your SaaS users aren’t all human. Explore how machine identities drive automation and create unseen risks across business applications.

2025年10月，Meta、微软、亚马逊接连把阿里通义千问、百度文心、月之暗面塞进自家云货架，价格直接砍到GPT的三分之一。

但是，让这样的天平始终保持平衡，是一个很理想化的乌托邦状态。而这个乌托邦状态的前提是，Meta的AI模型水平一直是要保持最领先的，或者说，至少是在开源赛道最领先，且不落后闭源模型太多的。